Here is a twist on the AI fake-citation saga. Harvard lecturer Duncan Levin, a former prosecutor, says new court rules asking lawyers to certify that a human checked AI-drafted filings could create a fresh problem. If a judge later asks whether you actually did the checking, the honest answer could be evidence against you.
Levin is careful about it. Embarrassment, malpractice exposure and sanctions alone do not trigger the Fifth Amendment. A careless miss is just careless. Real criminal risk needs more, like a knowingly false statement or a made-up explanation. Still, a lawyer at the podium deciding whether to invoke the privilege is a bad look for any client. GCs supervising litigation should ask the AI question first, before the judge does.
“Imagine that a brief contains fictitious cases. The court has reason to believe the lawyer failed to conduct a required verification, made an inaccurate representation about the use of AI, or tried to conceal what happened.”
Read: When a judge asks a lawyer about AI, can they plead the Fifth? at Harvard Law Today
